Improved spatial delineation of streambed properties and water fluxes using distributed temperature sensing

Abstract

A new method was developed for analysing and delineating streambed water fluxes, flow conditions and hydraulic properties using coiled fibre-optic distributed temperature sensing or closely spaced discrete temperature sensors. This method allows for a thorough treatment of the spatial information embedded in temperature data by creating a matrix visualization of all possible sensor pairs. Application of the method to a 5-day field dataset reveals the complexity of shallow streambed thermal regimes. To understand how velocity estimates are affected by violations of assumptions of one-dimensional, saturated, homogeneous flow and to aid in the interpretation of field observations, the method was also applied to temperature data generated by numerical models of common field conditions: horizontal layering, presence of lateral flow and variable streambed saturation. The results show that each condition creates a distinct signature visible in the triangular matrices. The matrices are used to perform a comparison of the behaviour of one-dimensional analytical heat-tracing models. The results show that the amplitude ratio-based method of velocity calculation leads to the most reliable estimates. The minimum sensor spacing required to obtain reliable velocity estimates with discrete sensors is also investigated using field data. The developed method will aid future heat-tracing studies by providing a technique for visualizing and comparing results from fibre-optic distributed temperature sensing installations and testing the robustness of analytical heat-tracing models.

Investigation of the thermal regime and subsurface properties of a tidally affected, variably saturated streambed

“…The temperature probe array data shows the temperature at six discrete depths at the upper and lower locations, whereas the coiled FO-DTS shows a near-continuum temperature-depth profile at these and an intermediate location. The temperature profiles illustrate relatively abrupt changes in temperature compared to other published time series (e.g., Rau et al, 2010;Vogt et al, 2012;Halloran, Roshan, Rau, Andersen, & Acworth, 2016c) due to the additional effect of the rising and falling ocean tides.…”

“…Temperature profiles were also recorded by coiled FO‐DTS (e.g., Vogt et al, ; Halloran et al, ; Figure c,d) with a vertical resolution of ∼8.7 mm. The coiled sections of each of the three coiled installations were ∼100 cm in height.…”

“…A pipe with nonrecoverable pointed drive head was used to create a guide bore for the FO‐DTS coils during installation. Lengths of fibre‐optic cable (∼ 25 m each) connecting the Sensornet Oryx FO‐DTS acquisition unit and the coils to one another were placed in a continuously agitated and cooled calibration bath with an accurate temperature reference to provide a two‐point calibration for each coil (Halloran et al ). The temperature probe arrays logged regularly at 5 min intervals and the FO‐DTS recorded at 10 min intervals with the notable presence of multiple gaps of 15–200 min duration due to intermittent cuts to the power generator.…”

“…The use of heat-as-a-tracer in surface water and groundwater investigations has seen a large amount of interest over the past decade (refer to following reviews of various aspects of heat-tracing: Anderson, 2005;Constantz, 2008;Rau, Andersen, McCallum, Roshan, & Acworth, 2014;Irvine et al, 2016;Halloran et al, 2016b), and some promising investigations using heat tracing techniques in the variably saturated zone have recently been undertaken (Béhaegel, Sailhac, & Marquis, 2007;Sayde et al, 2010;Krzeminska et al, 2012;Ciocca, Lunati, van de Giesen, & Parlange, 2012;Bechkit, Flageul, Guerin, & Tabbagh, 2014;Falocchi, Barontini, & Ranzi, 2015;Halloran, Roshan, Rau, Andersen, & Acworth, 2016c). For most passive methods that use natural temperature fluctuations, temporal resolution will generally be limited to 1 day when using the dominant diurnal signal.…”

Using Diurnal Temperature Signals to Infer Vertical Groundwater‐Surface Water Exchange

“…In cases where q values are relatively stable and are estimated at multiple depths, plots of mean (or median) q with depth can be a useful method to display results (Figure 7). In addition to Figure 7, examples of q-depth plots can be found in, Irvine et al (2015aIrvine et al ( , 2015b, Birkel et al (2016) and Halloran et al (2016). In cases where many temperature sensors with depth are available (e.g., when HRTS are used) it may be useful to present both temperature data (e.g., Figure 8a) and the calculated q values (e.g., Figure 8b) using surface plots with time on the horizontal axis, depth on the vertical axis, and the parameter of interest as the color.…”